The following message body content is getting around our Bayes filter (I added '-----' to the embedded domain name):
shank stalemate allentown budge dwelt villa annette surround euphemist consent discuss computation eldest psychotherapeutic dewdrop harpy cavil claret frame collet
This is a common spammer tactic. They add random words to try to bypass Bayesian filtering. This seems to be one of the dumber spammers, as they appear to be choosing random words rather than "good" words (ones more likely to be seen in legitimate mail than spam). Although "discuss" may be what put it over the edge -- I'm guessing that is significantly more common in legitimate mail than spam.
The non-related words are added to get around the Bayes filter (and are obviously successful). Can anyone recommend a good method to combat this? Should this message (and, consequently the added words) be processed by 'antispamseeder.exe' with the '/spam' switch, or is it best to remove the words before processing by 'antispamseeder.exe', or is it best to not process the message at all? Anyone have a good standards of practice for this?
The problem is that even if you put it through the antispamseeder, the next time they send a spam it will have different words. This is probably a matter of fine-tuning the Bayesian algorithm to better deal with these words.
-Scott
---
Declude JunkMail: The advanced anti-spam solution for IMail mailservers.
Declude Virus: Catches known viruses and is the leader in mailserver vulnerability detection.
Find out what you've been missing: Ask about our free 30-day evaluation.
--- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
